From M.A. Bozarth (1990). Drug addiction as a psychobiological process. In D.M. Warburton (Ed.), Addiction controversies (pp. 112-134 + refs). London: Harwood 

Academic Publishers. 
Drug Addiction as a 
Psychobiological Process
Michael A. Bozarth
Department of Psychology
State University of New York at Buffalo
Buffalo, New York 14260 U.S.A.
This chapter addresses the etiology of drug addiction. The emphasis is on biological mechanisms underlying addiction, although some other factors influencing drug addiction will also be discussed. The presentation is limited primarily to psychomotor stimulants (e.g., amphetamine, cocaine) and opiates (e.g., heroin, morphine) for two reasons. First, considerable knowledge has been gained during the past 15 years regarding the neurobiological mechanisms mediating their addictive properties. Second, these two pharmacological classes represent the best examples of potent addictive drugs, and the elucidation of their addiction potential can provide a framework for understanding abuse and addiction to other psychotropic agents.
Some psychologists and sociologists assert that animal studies do not model the important psychological variables governing drug addiction. They suggest that psychological processes critical in the etiology of addiction cannot be studied in animal models and/or that environmental influences important in producing an addiction cannot be duplicated in animal studies. This position is generally untenable, and animal models have been developed that accurately represent the primary processes involved in drug addiction. Support for the validity of these animal models will emanate from an understanding of the characteristics and the neural basis of drug addiction summarized in the following sections.
The arguments presented in the chapter are tenable, but they represent only one of several perspectives used in studying addiction. The terminology and even some aspects of the empirical data are the topics of scientific debate. The objective of this chapter is not to provide a balanced presentation of controversial issues, but rather to develop a unifying framework for understanding the psychobiological basis of addiction.
Concept of Addiction
Before proceeding with an examination of the mechanisms underlying drug addiction, it is necessary to define the term addiction and to examine the main characteristics of drug addiction. Delineation of the salient attributes of addiction helps to establish the criteria that must be fulfilled in a valid animal model and helps to determine what biological processes are relevant to the etiology of addiction.
Issue of Terminology
Drug addiction refers to a situation where drug procurement and administration appear to govern the organism’s behavior, and where the drug seems to dominate the organism’s motivational hierarchy. Jaffe (1975) has described addiction as "a behavioral pattern of compulsive drug use, characterized by overwhelming involvement with the use of a drug, the securing of its supply, and a high tendency to relapse after withdrawal [abstinence] (p. 285)." This definition follows the general lexical usage of the term and is consistent with the word’s etymology (see Bozarth 1987a).Drug addiction is defined behaviorally. It carries no connotations regarding the drug’s potential adverse effects, the social acceptability of drug usage, or the physiological consequences of chronic drug administration (Jaffe 1975). This latter point is especially important because some investigators have mistakenly used the term addiction to describe the development of physical dependence (see Bozarth 1987a, 1989; Jaffe 1975). Although drug addiction frequently has adverse medical consequences, it is usually associated with strong social disapproval, and it is sometimes accompanied by the development of physical dependence, these factors do not define addiction nor are they invariably associated with it. Drug addiction is an extreme case of compulsive drug use associated with strong motivational effects of the drug. Nature of Addiction
Initial drug use can be motivated by a number of factors. Curiosity about the drug’s effects, peer pressure, or psychodynamic processes can all provide sufficient motivation for experimental or circumstantial drug use. If the drug is taken repeatedly, a period of casual drug use often develops. Further use of the drug associated with more frequent drug administration, the use of higher drug dosages, and/or the use of more effective routes of administration (e.g., switching from intranasal to intravenous cocaine use) can lead to intensive patterns of drug use. Continued, more sustained drug use can then produce compulsive drug use where the substance has strong motivational properties and appears to govern much of the individual’s behavior. The most extreme case of drug use is the final progression to addiction. Drug use is viewed as a continuum, progressing from casual use to addiction (see Jaffe 1975); the drug assumes increasing control of the individual’s behavior as the pattern of drug use approaches addiction. Jaffe (1975) suggests that addiction is an extreme case of drug use that is not qualitatively different, but rather quantitatively different, from compulsive drug use. The failure to clearly distinguish between compulsive drug use and addiction appears to produce ambiguity and suggests a weakness in Jaffe’s (1975) definitions of these terms. Further consideration, however, reveals that an important inference can be made regarding the nature of addiction.
With this view—drug addiction representing the extreme point on a continuum progressing from casual drug use—drug addiction does not represent a special situation, but rather an extreme case of behavioral control. The only change is in the drug’s motivational strength and its disruption of the individual’s normal motivational hierarchy. (This latter effect has been termed motivational toxicity. See Wise and Bozarth 1985, for a discussion; see also Bozarth 1989 and Johanson et al. 1987). This represents a quantitative increase in the control of the individual’s behavior and not a qualitative shift in that behavior. With this perspective, addiction is an exaggerated form of normal behavior, similar to other types of psychopathology that represent extreme forms of exaggerated (compulsive) behavior. The distinguishing feature is the extreme motivational strength involving otherwise normal behavioral mechanisms. Therefore, it is a fundamental mistake to assume that addiction is a special case of behavioral control.

 Acquisition and Maintenance Phases
Drug addiction is frequently divided into two phases—acquisition and maintenance. This conceptual partition acknowledges that different factors may be involved in these two phases and that different degrees of drug-taking behavior are associated with these phases. The progression from the acquisition phase to the maintenance phase of addiction is not a quantal change, but rather it represents a shift in the importance of various factors that control the organism’s behavior along with an increase in the motivational strength of the drug-taking behavior. A brief example illustrates the utility of considering addiction as a two stage process.Prior to the first experience with a drug, the direct rewarding effects of drug administration are largely irrelevant in governing the individual’s behavior, except of course in that expectancies are developed from social interactions (e.g., media exposure, conversations with experienced users). Initiation of drug-taking behavior is governed by intrapersonal and sociological variables such as curiosity about the drug’s effects or peer pressure to try the drug. After initial exposure to the drug, pharmacological variables are relevant and will influence subsequent drug-taking behavior. Intrapersonal and sociological factors are probably still important in determining continued drug use, but they are less significant as the potent rewarding effects are repeatedly experienced. At some point there is a shift in control from intrapersonal/sociological to pharmacological factors in governing drug-taking behavior. This is concomitant with a marked increase in the motivational strength of the drug and with a progression from casual to compulsive drug use and ultimately to drug addiction. This may occur very rapidly for some drugs such as heroin or free-base cocaine and much more slowly for other drugs such as alcohol. The division of addiction into two separate phases does not presume that different mechanisms are involved in each phase. Rather, the demarcation acknowledges the possibility of different mechanisms but more importantly emphasizes differences in the motivational strength between the acquisition and maintenance of addictive behavior. As will be described later in the chapter, the same psychobiological process underlies both phases but additional variables are important in the acquisition of addiction. These other variables lose much of their influence as the addiction fully develops and as it becomes increasingly under control of basic pharmacological mechanisms.
Individual vs. Unitary Theories
A primary issue in considering the etiology of drug addiction is whether addiction to various drugs represents different processes, each specific to a particular drug type (i.e., individual theories), or whether some general mechanism underlies addiction to different pharmacological classes of drugs (i.e., unitary theory). A more extreme variation of the multiple theory approach might assert that the cause of addiction to even a single drug varies with each individual, thus necessitating unique theories for every case of addiction. In this latter situation, the causal elements in addiction would emanate primarily from psychodynamic processes, and the addiction would be viewed as nothing more than a specific instance of psychopathology. Treatment approaches used for other types of psychopathology would be appropriate, and no specialized procedures for treating addiction qua addiction would be necessary. This position has not gained popularity nor is it tenable as evidenced by the general failure of psychoanalytical and traditional psychotherapeutic methods to effectively treat drug addiction.
The possibility that addiction to different drugs involves a common mechanism has attracted many investigators, although most researchers confine their work to a single drug class. Attempts to identify underlying mechanisms common to various drug addictions do not presume that addictions to all classes of drugs are identical; there are obvious differences among addictions to different drugs, and even individual cases involving the same drug can display marked differences. However, certain elements of addiction seem to be shared across distinctively different pharmacological classes, and these similarities provide the impetus for developing unifying theories of addiction.
The unifying theory orientation suggests a somewhat different approach to studying addiction than does the individual theories orientation. First, drugs that produce the strongest addiction might be studied initially—the best examples of drug addiction should provide the best vehicle for identifying the underlying mechanisms. Drugs with weaker addictive properties would be examined after the relevant psychobiological processes have been delineated for drugs producing a rapid and profound addiction. Second, the commonalties among these addictive drugs should be identified and examined, and the differences should be presumed initially to have little importance in determining their addictive properties. The fact that one drug class produces signs of general behavioral stimulation and another drug class produces general behavioral sedation might be attributed to "side effects" of these drugs and not deemed important in understanding their addictive properties. Third, individual theories of addiction would be developed for different drugs only as conclusive evidence showed that the more general theory was not adequate. This principle of parsimony has been useful in resolving other, seemingly complicated phenomena into simpler conceptualizations. Animal Models
Several animal models of human drug addiction have been studied. Some involve the interaction of addictive drugs with electrical activation of brain reward pathways, while others have studied the various behavioral and physiological effects of drugs (see Bozarth 1987ab). The most popular methods have focused on the ability of drugs to directly control the animal’s behavior. This approach is consistent with the behavioral definition of addiction, and it has the strongest face validity of any animal model used to study human drug addiction. Using traditional operant psychology techniques, laboratory animals can be trained to self-administer many psychotropic drugs. Although animals will self-administer drugs by various routes of administration (e.g., oral, intragastric, intracranial), the intravenous self-administration method has gained the most widespread acceptance. Animals are surgically prepared with intravenous catheters and are tested for voluntary drug self-administration using traditional operant techniques (see Yokel 1987). Typically, the subjects are tested in an operant chamber containing a lever; depressing the lever automatically delivers drug through an intravenous catheter. Experimental procedures have been developed that permit testing of unrestrained, freely moving subjects. With this technique, normal animal behavior (e.g., grooming, feeding and drinking) can be studied concurrently with intravenous drug self-administration.
Addictive drugs control behavior in a manner similar to conventional reinforcers (e.g., food and water) when drug administration is made contingent upon lever pressing (Johanson 1978; Spealman and Goldberg 1979; see also Fischman and Schuster 1978). Most drugs that are addictive in humans are readily self-administered by laboratory animals, and drugs that are not addictive in humans are generally not self-administered by animals (Deneau et al. 1969; Griffiths and Balster 1979; Griffiths et al. 1979a; Weeks and Collins 1987; Yokel 1987). Procedures used to study intravenous drug self-administration in laboratory animals have also been applied to studying drug self-administration in humans (see Henningfield et al. 1987; Mello and Mendelson 1987).
Approximately 80% of the animals tested for intravenous cocaine or heroin self-administration learn to self-administer drug under standard laboratory conditions (see Bozarth 1989). No special training procedures or pre-existing conditions (e.g., food deprivation) are necessary for these drugs to serve as rewards in this experimental paradigm. If operant shaping techniques are used, this number approaches 100%. Some animals learn within several hours of exposure to the testing procedure, while others may require two or three weeks of exposure for several hours each day before reliable patterns of drug self-administration emerge. Animals tested under limited access conditions (viz., no limitations on the amount of drug administered per hour, but subjects can only self-administer drug for a limited number of hours each day, e.g., 2 to 12 hours daily) maintain good general health and show little or no disruption of food and water intake. Limited access testing is the procedure used most often in intravenous self-administration studies, and it is associated with low subject morbidity and attrition. Testing cocaine under unlimited access conditions (i.e., continuous testing 24 hours per day) is accompanied by an extremely high subject mortality (90% subject loss within 30 days; Bozarth and Wise 1985), and it produces a rapid deterioration in the animal’s health. For this reason, the unlimited access procedure has been used very infrequently, and all further discussion of this method will be restricted to limited access conditions. Animals tested for intravenous psychomotor stimulant or opiate self-administration quickly develop stable patterns of drug intake, where the average hourly drug intake is consistent both within and between experimental sessions. The effect of changing the amount of drug administered with each injection (i.e., unit dose) is predictable, and the substitution of saline for reinforcing drug produces a rapid extinction of lever-pressing behavior. The intravenous self-administration procedure has been used extensively to study the behavior maintained by drugs serving as reinforcers and to study the neural basis of drug reward.
Neural Basis of Drug Reward
The majority of research investigating the neural mechanisms of motivation and reward has been conducted using laboratory animals. Although most scientists see no difficulty in generalizing from these studies to human neurobiology, brief mention of the applicability of these data is warranted. First and foremost is the recognition that there are obvious anatomical and physiological differences, but the primary difference between laboratory rats (the most commonly used species) and phylogenetically higher mammals is in cortical development. These higher brain centers are involved primarily in cognitive processes such as learning and memory, in speech, and in fine motor control. The basic motivational substrates across mammalian species are probably very similar. The limited neurophysiological and pharmacological investigations that have been conducted in humans seem to confirm this notion of similar brain reward pathways (e.g., Heath 1964). Second is the acknowledgement that motivational differences do exist, but that the most important difference between human and infrahuman animals probably involves cognitive influences on these motivational mechanisms. These influences cannot be fully studied in animal models, but they probably exert their primary influence on initial drug-taking behavior and have much less influence once intensive patterns of drug taking have developed.
Brain dopamine systems have been the focus of considerable attention in behavioral neurobiology. In particular, the ventral tegmental dopamine system appears to have an important role in motivated behavior (see Bozarth 1987c) and in some types of psychopathology. This dopamine system has its cell bodies located in the ventral tegmental area and sends its axonal projections to several brain regions (see Lindvall and Bjorklund 1974; Ungerstedt 1971a), most notably the nucleus accumbens (see Figure 2). It receives neural inputs from many diverse brain sites and modulates neural activity in cortical and limbic areas.


Psychomotor Stimulant Reward
The component of neural transmission generally most sensitive to pharmacological manipulations is synaptic activity. Neurotransmitters are released following the arrival of an action potential at the presynaptic terminal and rapidly diffuse across the synaptic cleft to postsynaptic target cells. Once bound to their receptors, they can either facilitate or inhibit neural activity in these target neurons. Psychomotor stimulants strongly affect catecholaminergic synaptic transmission (viz., neurons releasing dopamine or norepinephrine). Cocaine blocks the inactivation of dopamine by inhibiting its presynaptic reuptake (Heikkila et al. 1975) thereby increasing the effect of synaptically released dopamine; amphetamine blocks dopamine reuptake and also inhibits its degradation by monoamine oxidase (Axelrod 1970; Carlsson 1970). Both actions produce a potent enhancement of dopaminergic neurotransmission. Other neurotransmitter systems are also affected by psychomotor stimulants (e.g., noradrenergic, serotonergic), but several studies have shown that enhancement of dopaminergic neurotransmission is critically involved in the rewarding action of these drugs.
Neuroleptic drugs—which block dopamine receptors—disrupt the intravenous self-administration of psychomotor stimulants, while drugs blocking noradrenergic receptors are ineffective (de Wit and Wise 1977; Yokel and Wise 1975, 1976; see also Yokel 1987). Lesions of the dopaminergic terminal field in the nucleus accumbens attenuate psychomotor stimulant self-administration (Lyness et al. 1979; Roberts et al. 1977, 1980; see also Roberts and Zito 1987), as do lesions of the dopamine-containing cell bodies in the ventral tegmental area (Roberts and Koob, 1982). These studies have used a selective neurotoxin that destroys only dopamine neurons and has no appreciable effect on the other neurons found in those areas. Self-administration procedures have been adapted so that animals can self-administer drug directly into restricted brain areas (see Bozarth 1983, 1987d). Studies using this intracranial self-administration technique have shown that amphetamine (Hoebel et al. 1983) or dopamine (Dworkin et al. 1986) injections administered directly into the nucleus accumbens are rewarding. These lines of evidence have firmly established a role of the ventral tegmental dopamine system in psychomotor stimulant reward.Opiate Reward
Opiates do not appear to affect dopaminergic synaptic activity directly but do stimulate dopamine neurons by an action at the cell body region in the ventral tegmentum. Following opiate administration the neural activity of these dopamine neurons is increased (Gysling and Wang 1983; Matthews and German 1984). Action potentials generated at the cell body region are conducted along the axon to the synaptic terminals in the nucleus accumbens (see Figure 2). There they produce an impulse-coupled release of dopamine. The increased cell firing rates in the ventral tegmentum lead to an increased dopamine release in the nucleus accumbens (Di Chiara and Imperato 1988; Westerink 1978; Wood 1983). Both the action of opiates in the cell body region (enhancing dopamine cell firing rates) and the action of psychomotor stimulants in the terminal region (enhancing dopaminergic synaptic activity) produce a net increase in dopaminergic neurotransmission in the nucleus accumbens. Different neural elements are involved, but an important neural action is shared by both classes of drugs.
Dopamine-depleting lesions of the ventral tegmental area disrupt the acquisition of intravenous heroin self-administration (Bozarth and Wise 1986). The effects of neuroleptics on opiate self-administration have been difficult to interpret (see Bozarth 1986; Wise 1987; cf. Ettenberg et al. 1982), but conditioning studies have shown that neuroleptics block opiate reward (Bozarth and Wise 1981a; Phillips et al. 1982). Animals will readily self-administer opiates directly into the ventral tegmental area (Bozarth and Wise 1981b; Van Ree and De Wied 1980), and the rewarding action of these injections has been confirmed using other behavior techniques (Bozarth and Wise 1982; Phillips and LePiane 1980). The anatomical zone where morphine infusions are rewarding corresponds closely to the location of the dopamine-containing cell bodies in the ventral tegmental area (see Bozarth 1987e). Infusions of morphine directly into the ventral tegmentum do not produce physical dependence, while morphine infusions into another brain site that does produce physical dependence (i.e., the periaqueductal gray region; see Figure 2) are not rewarding (Bozarth and Wise 1984). This neuroanatomical dissociation of reward and physical dependence shows that opiates can be rewarding without the development of physical dependence.
The interpretation of research identifying the neural basis of opiate reward has been somewhat controversial, but considerable data suggest that opiates can activate the same brain reward system as that mediating reward from psychomotor stimulants. Direct support for this hypothesis comes from a study showing that ventral tegmental morphine injections can partially substitute for intravenous cocaine injections (Bozarth and Wise 1986). This would be expected if the same brain reward system is critically involved in the rewarding actions of these two classes of drugs. In addition, chronic opiate administration may evoke other reward processes that are not shared with psychomotor stimulants, but these processes are not important in the initial rewarding action of opiates.
Other Rewards
Other drugs may activate the ventral tegmental dopamine system; alcohol and nicotine have been shown to increase dopamine release in the nucleus accumbens (Di Chiara and Imperato 1988). The importance of this effect for the rewarding actions of these compounds has not been systematically evaluated, but it seems likely that at least part of their rewarding properties may evolve from an action on this reward system. Furthermore, the rewarding effect of electrical brain stimulation (at least from some electrode sites) appears to involve dopaminergic neurotransmission (Fibiger 1978; Fibiger and Phillips 1979; Wise 1978; see also Bozarth 1987c), and the regulation of food and water intake has an important dopaminergic component (see Ungerstedt 1971b; Wise 1982). These data suggest that brain dopamine systems—in particular the ventral tegmental dopamine system—may provide a general motivational function, and this hypothesis is consistent with the notion that addictive drugs derive their rewarding effects by pharmacologically activating brain reward systems involved in governing normal behavior. (See Bozarth 1986, 1987c; Wise and Bozarth 1984, 1987. For earlier discussions of the interaction of addictive drugs with brain reward systems, see Broekkamp 1976; Esposito and Kornetsky 1978; Kornetsky et al. 1979; Reid and Bozarth 1978).Pre-eminent Role of Animal Studies
There are numerous findings from animal studies that have important implications for understanding human drug addiction. Some of these findings contradict commonly held notions about drug addiction and others resolve issues where clinical research has been indecisive. Animal research in several important areas can help direct future clinical research and can prompt a reinterpretation of some clinical studies. This is the converse of the usual situation where animal studies are considered inadequate if they fail to meet expectations generated by clinical studies. Animal studies have the advantages (i) of not being limited by the ethical constraints imposed on human clinical research and (ii) of having a subject population where the important variables can be adequately controlled. Four examples will be used to illustrate the pre-eminence of animal research.
First, animal studies have clearly shown that pre-existing conditions are not necessary for drugs to be rewarding. Psychopathology, stress, and other intrapersonal conditions may influence drug-taking behavior, but animal research has shown that none of these factors are necessary conditions for a drug to exert its potent ability to control behavior. Sociological variables (e.g., deviant or rebellious behavior, peer pressure, modeling) may also influence drug-taking behavior but are obviously not essential for drugs to serve as reinforcers. Mere exposure to the drug is sufficient to motivate subsequent drug-taking behavior.
Second, the self-administration of addictive drugs by laboratory animals supports the notion that drugs act as universal reinforcers. Human-specific attributes are not necessary for drug reinforcement to occur. The factors governing drug-taking behavior are not unique to humans; they involve biobehavioral processes shared across mammalian species. Indeed, drug addiction might be considered a phylogenetically primitive behavior. The brain systems mediating the addictive properties of drugs evolved early and have a central role in promoting survival of the organism.
Third, neural mechanisms involved in opiate and psychomotor stimulant reward have been identified. An important component of reward from these drugs involves the activation of a common neural substrate, although additional brain systems may also be involved. This suggests a commonalty between these distinctively different pharmacological classes. This shared action on a brain reward system has long been obscured in animal and human studies noting the marked differences in the general effects produced by these drugs. Identification of the brain mechanisms underlying the rewarding actions of these drugs has relegated the prominent differences to the status of "side effects." Just as the cardiovascular effects of psychomotor stimulants are unlikely contributors to their addictive properties, the analgesic effects of opiates are probably not important in controlling opiate addiction. (Physical dependence and withdrawal from chronic opiate administration remains an unsubstantiated, but potentially significant, factor in long-term opiate addiction; see Bozarth 1988; Bozarth and Wise 1983.) Human subjects exposed to the complex stimulus properties produced by drug administration are likely to attend to and report the most salient interoceptive cues. Many of these cues, frequently related to general central nervous system stimulation or depression, may overshadow and mask the similar subjective states produced by various drugs unless adequate subjective-effects measures are used. Some early investigators recognized the importance of subjective effects such as mood elevation and euphoria (Eddy 1973; Eddy et al. 1957; McAuliffe and Gordon 1974), but most seem to have been preoccupied with the drugs’ general central nervous system effects which can differ markedly for various classes of addictive drugs. The Addiction Research Center Inventory, an empirically derived test designed to measure the subjective effects of addictive drugs, detects the mood-elevating effects of psychomotor stimulants and opiates on the same scale (see Haertzen and Hickey 1987). On the other hand, the subjective-effects of these two drug classes can be easily distinguished. This is not surprising considering that ex-addicts report a preference for heroin over morphine (Martin and Fraser 1961), even though heroin is rapidly converted to morphine after entering the brain (Jaffe and Martin 1975). This drug preference is probably related to pharmacokinetic differences in these two compounds which may produce differences in their interoceptive cues. Fourth, clinical research often suffers from strong subject biases. Response-demand characteristics can have a large influence on the subject’s responses, not only affecting the intensity but also the direction of responses. The power of these demand characteristics is perhaps best illustrated by hypnotic phenomenon, where seemingly supranormal behaviors can be elicited (e.g., Barber 1972) and where the subject’s uncertainty about events can be supplanted with absolute confidence (e.g., Laurence and Perry 1983; Perry and Laurence 1983). The subject’s behavior may also be influenced by the consequences of his responses. Consider, for example, the methadone maintenance patient who wishes to have his methadone dose increased to experience mood elevation or euphoria. If the patient tells the physician that he wishes to recapture the pleasant subjective state produced by the opiate, the physician is unlikely to prescribe an increased dosage. If, on the other hand, the patient reports intense subjective discomfort related to opiate withdrawal reactions, this may evoke an empathetic increase in the methadone dosage. Similarly, expressions about craving and desire for the drug are less likely to receive favorable support from the physician than complaints about the physical symptoms of withdrawal. Craving and desire are "psychological" attributes and frequently viewed as ‘under the person’s control;’ withdrawal discomfort is attributed to "physiological" processes that are not directly controlled by the patient. Much like any physical illness, physiological withdrawal reactions are considered nonvolitional and therefore foreign to the individual’s concept of ‘self.’ The response-demand characteristics of this situation can evoke exaggerated emphasis on withdrawal reactions and obscure the importance of other factors such as drug craving. (It is important to note that a person may mistakenly attribute his drug-taking behavior to these factors even in the absence of significant situational demand characteristics. This false attribution rationalizes the individual’s tendency to consider the addiction nonvolitional, ‘out of his control.’ Recent clinical trends have shifted toward the recognition of drug-induced alterations in desires and their fundamental role in addiction.) Animal studies have permitted an unmasking of important effects shared by different classes of drugs and have provided much of the impetus for developing unifying theories of addiction. These studies have a seemingly pre-eminent role in directing research with humans, refocusing attention on central issues in drug addiction and suggesting a reinterpretation of some human studies. Dissonance between animal and human studies probably indicates that the human studies are influenced by additional factors. These factors are most likely to be factors important in the etiology of addiction that are not duplicated in animal studies or factors related to subject/experimenter biases resulting from inadequately designed clinical research. Inadequately designed animal research could also produce dissonance, but this is less likely because the relevant variables are more easily controlled. The burden of scrutinizing the experimental design and results falls most heavily on the clinical studies, and animal studies have attained a pre-eminent role in delineating addictive processes.
Relevance of Motivational Theory
The nature of drug addiction places its study firmly in the realm of motivational psychology. Many of the experimental methods that have been developed to study conventional rewards (e.g., food, water, sex) can be applied to the study of addiction, and the conceptual advances made in motivational theory can be used to guide the study of addictive behavior. The consideration of addiction as simply exaggerated/excessive behavior (viz., a pathological manifestation of normal behavioral processes) prompts fitting addiction into the framework provided by general motivational theory.
Importance of Considering Motivational Strength
The defining feature of addiction is its potent control of behavior—the motivation to obtain and self-administer the drug is extremely strong. Several behavioral measures have been used to determine motivational strength; the primary measures are response latency, response frequency, and response vigor. Although the first two measures have been used to quantify motivational strength, response vigor measures are studied most frequently in classic animal motivation studies (e.g., see Bolles 1967, 1975; Hull 1943, 1951). Response vigor can be subdivided into three general types of measures: resistance to extinction, work output to obtain reward, and magnitude of aversive stimuli necessary to suppress responding for the goal object. All of these variables increase as a monotonic function of motivational strength across a variety of conventional rewards (see Bolles 1975).Techniques are available to study these response vigor measures in animal models of addiction, but they have not been routinely used. However, two specific experimental methods have been used to measure the strength of drug-taking behavior, and they will be briefly described here. Both methods employ intravenous drug self-administration.
After drug self-administration has been established, substitution of drug vehicle for the reinforcing drug produces an extinction pattern similar to that produced by conventional rewards such as food and water—there is an initial increase in lever pressing followed by a cessation in responding. Noncontingent, experimenter-delivered priming injections of the drug can reinstate lever-pressing behavior on subsequent trials even though the lever-pressing fails to produce injections of the reinforcing drug. This technique may provide an animal model of relapse and may be related to the human subjective experience of craving. Indeed, the animal’s responding despite the response-contingent delivery of rewarding drug may represent a type of drug-seeking behavior. Priming with the rewarding drug reinstates behavior previously reinforced by drug injections, and this effect is conceptually related to the classic response vigor measure of resistance to extinction. (See Stewart and de Wit 1987, for a discussion of this method.)Another experimental method that measures response vigor in animals trained to self-administer drugs is a variation of traditional operant methodology. Animals can be readily trained to lever press several times for each drug injection. Once the behavioral response is firmly established, the number of lever presses required to produce a drug injection is progressively increased. The number of times an animal will lever press for a single reinforcing drug injection can be interpreted as reflecting the work output for the drug. This provides a measure of the drug’s motivational strength, and this technique has been successfully used by several laboratories (e.g., see Brady et al 1987; Roberts et al. 1989; Yanagita 1987).Both of these experimental procedures appear to measure the motivational strength of the drug reward. Their face validity is high, and increases in the amount of drug delivered during priming or during self-administration produce increases in these measures. Drugs addictive in humans (e.g., cocaine, heroin) have potent motivational properties as demonstrated by these techniques.
Despite the availability of these and other potentially valuable methods for determining the motivational strength of various drugs, they have seldom been employed. This is unfortunate because the case for addiction can only be established by demonstrating the extremely strong motivational properties of the drug that are inherent in the definition of addiction. Nonetheless, there is an excellent correspondence between drugs that are self-administered by laboratory animals and drugs that are clinically judged to be addictive in humans. Despite this concordance, however, it is erroneous to consider a drug addictive just because it is self-administered by animals just as it is erroneous to diagnose addiction on the simple observation that a substance is taken by humans. To establish that a drug is addictive, it is essential to show that the substance has the strong motivational properties necessary to produce the level of compulsive drug-taking behavior that defines addiction. Simple demonstration of drug reward is insufficient, because substances may activate brain mechanisms involved in motivation and reward without activating them in a manner that produces the extreme, exaggerated behavior termed addiction.
An Incentive Motivational Model
There are two contrasting positions regarding basic motivational mechanisms that will be considered. Drive-reduction theory asserts that organisms are motivated by drives which ‘push’ the animal toward the goal object; the primary motivation is to reduce the drive. Incentive motivational theory asserts that organisms are motivated by incentives (viz., attraction to the goal object) which ‘pull’ the animal toward the goal object; the primary motivation is the expectancy of reward. With drive-reduction theory, it is the termination of some condition that motivates the organism (e.g., the reduction of hunger); with incentive motivation theory, it is the elicitation of some condition that motivates the organism (e.g., the anticipation of eating when hungry). The following study illustrates these two contrasting theories of motivation. Feeding can be produced by electrical stimulation of the lateral hypothalamic area in food-satiated animals. This stimulation-induced feeding probably involves brain systems mediating the normal control of eating. (See Wise 1974, for a discussion of stimulation-induced feeding.) Because feeding begins immediately upon activation of the electrical current and stops soon after its termination, Mendelson (1966) was able to use electrical stimulation to produce ‘hunger’ in selected parts of a T-maze. This permitted an analysis of the "Role of Hunger in T-Maze Learning for Food by Rats." The study’s objective was to determine where in the T-maze the animal must be ‘hungry’ to select the goal box containing food. The motivational condition produced by the electrical stimulation (which is functionally equivalent to natural hunger; Wise 1974) was elicited and terminated in various parts of the experimental apparatus. Mendelson (1966) first showed that food-satiated animals would run to the compartment containing food when electrical stimulation was constantly applied throughout the T-maze (i.e., in the start box, runway, choice point, and goal box). Stimulation terminating upon the animal’s entry into the goal box was not sufficient to maintain the behavior, but stimulation activated only upon entry into the goal box containing food was effective. The seemingly surprising aspect of this study (at least for drive-reduction theorists) was that termination of stimulation-induced ‘hunger’ did not maintain responding. In fact, the motivational condition produced by the stimulation was not even necessary at the choice point where the animal selected the compartment containing the food; it only had to be present in the goal box containing food. Mendelson’s (1966) interpretation was consistent with current incentive motivational theory: the experience of having eaten on previous trials (reward) was both necessary and sufficient for this instrumental response. (The demonstration of instrumental responding for food in the absence of drive has been independently replicated; Streather et al. 1982). Numerous studies have shown similar effects, and attempts to demonstrate behavior maintained by drive reduction (e.g., delivery of intravenous water to water deprived rats; Corbit 1965) have been largely unsuccessful.A detailed critique of motivational theory is beyond the scope of this paper, but drive-reduction theory is considered generally untenable by specialists in motivational theory (see Bindra, 1969, 1974, 1978; Bolles 1972, 1967, 1975; Toates 1981). There has been a shift to incentive motivational explanations—particularly for appetitively motivated behaviors—which provide a more adequate explanation of the empirical data as well as a more satisfactory theoretical integration (see Bindra 1969, 1974, 1978; Bolles 1972, 1967, 1975; Toates 1981). Drive-reduction theory probably best explains conditions of aversive motivation (i.e., negative reinforcement processes), where the animal is avoiding or escaping an aversive stimulus (see Spence 1956, 1960). Incentive motivational theory better explains conditions of appetitive motivation (i.e., positive reinforcement processes), where the animal is approaching some stimulus associated with reward. Drug addiction appears to be governed primarily by an incentive motivational process like the motivational properties of other appetitive stimuli. Incentive motivational theory focuses on associative conditioning like other learning theories, but the expectancy of reward is the primary motivator. It is a cognitive theory, where cognitive processes and anticipation figure prominently. Some versions of incentive motivational theory also provide a basis for understanding emotions (Bindra 1978), and a similarity with traditional hedonic theories (e.g., Pfaffman, 1960; Young 1959, 1961, 1966) is apparent. Motivational theorists have generally abandoned drive-reduction theory as a universal principle governing most behavior. Unfortunately, many other psychologists and physiologists are still preoccupied with the notion of homeostatic regulation of needs. Homeostatic regulation of various physiological processes (e.g., body temperature, cardiac output) has provided an excellent model for understanding many biological processes, but it has provided an inadequate foundation for explaining most behaviors. The failure of many scientists to abandon the notion of behavioral homeostasis has obscured a better understanding of motivated behavior.
A drive-like mechanism may influence drug-taking behavior in some conditions. Drugs that produce physical dependence and that when discontinued produce pronounced physical withdrawal reactions accompanied by subjective feelings of distress may have an additional mechanism for maintaining drug-taking behavior. Because drug administration can relieve the withdrawal distress associated with abstinence, the termination or avoidance of withdrawal discomfort might also influence drug-taking behavior. This negative reinforcement process would have the characteristics described by drive-reduction theory. Nonetheless, this potential negative reinforcement mechanism is not necessary for drugs to be reinforcing nor is it an adequate explanation for initiation or relapse to drug use. Indeed, the ability of such a negative reinforcement process to maintain drug-taking behavior has not been experimentally demonstrated.
Addiction can generally be considered an appetitively motivated behavior, governed by the incentive properties of the drug and related stimuli. The expectancy of reward (developed from previous drug-taking experience) and the subsequent pharmacological activation of brain reward pathways constitute a sufficient explanation of the late acquisition and maintenance phases of addiction. Other factors (e.g., intrapersonal, sociological) probably have a primary role during the early acquisition phase of addiction.
Addiction to Commonly Used Substances?
There are several widely used substances that are generally considered socially acceptable—alcohol, caffeine, and nicotine. Alcohol has well documented addictive properties, although there is some debate over what constitutes alcoholism and how much time is required to develop alcohol addiction. Of the remaining two substances, nicotine has come under increasing scrutiny as a potentially addictive substance. Indeed, several organizations and some scientists have suggested that nicotine is an extremely addictive drug (even the most addictive drug), and there has been recent legislative attempts to regulate its usage. Regulation of smoking is related to its reputed health hazards, but the persistence of smoking behavior has generated debate over the possible addictive properties of nicotine. Nicotine use can produce an elevation in mood (e.g., Henningfield et al. 1985), and some studies suggest that nicotine intake may be necessary for optimal functioning in chronic smokers. Intravenous nicotine self-administration has been demonstrated in humans (see Henningfield et al. 1987), and smoking behavior appears to be largely influenced by the nicotine content of the cigarettes (see Henningfield and Goldberg 1988). A few animal studies have reported intravenous nicotine self-administration, although difficulties have also been reported in obtaining reliable self-administration (see Balfour 1982; Dougherty et al. 1981; Henningfield and Goldberg 1983).
The argument suggesting that nicotine is an addictive substance appears to be supported by the following observations which reveal similarities with prototypical addictive drugs like cocaine and heroin: (i) human subjective evaluations showing it can elevate mood (e.g., Henningfield et al. 1985), (ii) intravenous self-administration studies showing that it can serve as a reinforcer in laboratory animals (see Henningfield and Goldberg 1983), and (iii) neurochemical studies showing apparent activation of the ventral tegmental reward system (e.g., Andersson et al. 1981; Di Chiara and Imperato 1988). Each line of evidence may seem sufficient to demonstrate that nicotine is addictive, but they all suffer from one common misconception. This misconception becomes obvious upon considering the fundamental nature of drug addiction.
None of these methods assess the motivational strength of nicotine administration. First, the subjective-effects measures are probably best considered qualitative measures, determining if various drugs have the ability to modify mood and affect in a manner similar to addictive drugs. Other manipulations may produce similar alterations in mood and affect, and these measures by themselves do not establish the addiction liability of a substance. Positive findings with these measures do indicate a potential addiction liability, but they fail to establish that the compound can control behavior to the degree necessary to produce a true addiction. Second, animal self-administration studies have shown that nicotine can serve as a reinforcer under some experimental conditions, but none of these studies have evaluated the strength of drug-taking behavior. In fact, nicotine self-administration in laboratory animals would appear much more difficult to establish than self-administration of psychomotor stimulants or opiates (see Griffiths et al. 1979b). This suggests that the reinforcing effect of nicotine in laboratory animals may be much weaker than that of prototypical addictive agents. (Even if strong motivational effects were demonstrated for nicotine, the difficulty in establishing nicotine self-administration suggests that it has a relatively low [compared with cocaine or heroin] addiction liability; see Bozarth 1989.) Third, activation of the ventral tegmental reward system does not necessarily lead to addiction, just as periodic activation of this system does not invariably produce an addiction. Normal behavior may be partially directed by the activity of this system (e.g., feeding, see Hamilton and Bozarth 1988; sexual behavior, see Pfaus et al. 1989), and the usual expression of these behaviors would not be described as constituting the degree of compulsiveness necessary to fulfill the definition of addiction. Drug addiction may require more than just simple activation of this reward system and may even involve neuroadaptive changes in this system (see Bozarth 1989; Dackis and Gold 1985; see also the chapter by White, this volume).
The above discussion is not meant to argue that nicotine is a nonaddictive drug. It is intended to refute the notion that nicotine has an unequivocally established addiction potential of equal magnitude to prototypical addictive agents; this is clearly a hasty conclusion that exceeds the available data base. Second, it serves to illustrate the principle that substances can share many effects with highly addictive drugs and not necessarily be addictive themselves. This latter point is most important in conceptualizing the very nature of addiction and in understanding the interaction of addictive substances with brain mechanisms subserving motivation and reward processes. With the perspective advocated by this chapter, drug addiction is viewed as an extension of normal behavioral processes, and the addiction potential of a drug is derived from its ability to activate brain mechanisms involved in the control of normal behavior. Drug addiction represents a case of extreme control exerted by a pharmacological substance that can disrupt the individual’s motivational hierarchy. "Psychological" vs. "Physiological" Processes
One point regarding the psychological and the physiological natures of behavior deserves special mention. Obviously, psychological events have some basis in brain physiology, but a strictly reductionistic approach to behavior frequently ignores important cognitive processes. Physiological processes affect/produce cognitive events and cognitive events affect/produce physiological processes. What is considered "psychological" and what is considered "physiological" in nature is largely determined by one’s perspective .Drugs can affect psychological events and this may be reflected as changes in desires and motivation. The initial rewarding effects of many drugs are probably experienced subjectively as an elevation in mood and affect (see Haertzen and Hickey 1987; Henningfield et al. 1987; McAuliffe and Gordon 1974). The subjective state produced by the drug, however, is clearly distinct from the ‘normal’ psychological feelings of ‘self.’ Repeated experience with an addictive drug may produce a breakdown in this distinction. Conditioning processes may elicit cognitions about the drug and its appetitive effects. These "psychological" events may be accompanied by the subjective experiences of desire and craving. As the addiction develops fully, changes in the individual’s motivational hierarchy ensue; the drug whose effect was sought only occasionally and whose intake was limited by intrapersonal and sociological factors begins to dominate the individual’s behavior. Other formerly potent motivators (e.g., food, sex, safety) lose their abilities to influence the individual’s behavior, and motivational toxicity usually becomes apparent. This progression from casual drug use to addiction results from the interaction of the drug with brain reward systems and from cognitive processes related to anticipation of the drug’s rewarding effects. Desire and craving can be elicited by physiological events. Through this process a drug’s pharmacological action can alter feelings of the ‘self’ and enter the realm experientially labeled ‘the mind.’ A desire for the drug can develop from repeatedly experiencing its rewarding effects; this desire is phenomenologically within the ‘self’ and not distinguishable as externally controlled behavior any more than feelings of hunger or thirst are considered under the control of external factors. What may not be immediately apparent is that cognitions and social interactions can also affect ‘physiological’ processes. Associative processes (e.g., exposure to stimuli related to drug taking) may elicit subtle activation of brain reward mechanisms. This, in turn, may produce a priming effect eliciting motivational arousal and intensifying the incentive value of the stimulus conditions associated with the drug. (An animal model of this phenomenon may be the conditioned place preference paradigm where animals approach environmental cues previously associated with drug reward; see Bozarth 1987f, White et al. 1987.) The subjective experience of craving may accompany this subtle activation of reward processes, and cognitive processes may further exacerbate physiological activation and craving. Cognitive techniques that disrupt this cycle may abate craving and diminish its subsequent effect on behavior (e.g., relapse to drug taking). A Simple Psychobiological Schema
Social, personality, and cognitive factors are very important in instigating and maintaining drug usage during the acquisition phase of addiction. Obviously, the pharmacological effects of a drug—no matter how powerful—cannot provide the impetus for initial drug use. It is likely that early drug use is largely governed by nonpharmacological factors, although the biological consequences of drug administration will quickly have some influence on subsequent drug usage. At some point during repeated drug use (very quickly for some compounds), the pharmacological actions of the drug usually predominate and the other factors influencing drug intake have less significance. The extreme case of drug use (i.e., addiction) is primarily under control of the pharmacological effects on brain motivation and reward mechanisms, and it is this phase of drug usage that has been the focus of this chapter. A simple psychobiological schema for conceptualizing the etiology of drug addiction is illustrated in Figure 3. It depicts three domains that govern addiction—intrapersonal, sociological, and pharmacological. During the acquisition phase, personality and environmental factors can play important roles in drug use. These are primarily within the intrapersonal and sociological domains, respectively, while circumstantial factors that can influence drug use fall within either domain. As drug usage intensifies, it becomes progressively under control of factors in the pharmacological domain. With this schema, addiction is seen mainly as a pharmacological process involving the interaction of the drug with brain mechanisms mediating motivation and reward. Other factors important in the genesis of drug addiction, however, include intrapersonal and sociological events, and they are acknowledged by the schema illustrated by the figure. The most obvious way that intrapersonal and sociological factors can affect the development of addiction is by influencing the degree of continued drug exposure. Intrapersonal and sociological factors may facilitate drug-taking behavior (e.g., rebellious tendencies and peer-pressure, respectively) or they may inhibit repeated drug use (e.g., fear of adverse medical consequences and legal/social sanctions, respectively), especially early during the acquisition phase. These factors do not directly affect the pharmacological reward produced by drug administration but merely modulate the continued exposure to the addictive drug. Another way that intrapersonal/sociological factors may influence drug-taking behavior during the acquisition phase is by affecting the rewarding action of the drug. For the first few weeks of testing, individually housed animals intravenously self-administer more heroin than socially housed animals; this effect seems to be limited to influencing the rate of acquisition because both groups of subjects learn to reliably self-administer heroin (Bozarth et al. 1989; cf. Alexander 1984). However, it does demonstrate that a social manipulation (isolation distress?) can influence the pharmacological reward produced by a drug.
The psychobiological approach to studying drug addiction emphasizes the importance of neural mechanisms that govern normal behavior. Addiction is not viewed as a unique condition nor does it differ significantly from other forms of compulsive behavior. Rather, it involves conventional motivational processes and is distinguished only by its extremely potent control of behavior. The rewarding drug effects involve primarily an appetitive motivational process best described by incentive motivational theory. Cognitive expectancies figure prominently in the individual’s behavior, and intrapersonal and social factors can significantly influence drug taking. Addiction, however, is most directly related to the drug’s pharmacological properties.